The present technology relates to an image processing device and an image processing method, and particularly to enabling a noise-removed image of excellent image quality without a loss of sharpness in edge parts to be generated in removing noise from an image.
In related art, an image processing device removes noise from an image when extracting edge information from the image and performing image processing, for example, because the extraction of the edge information is inevitably affected by a noise part when the noise part is included in the image. In addition, in noise removal, performing image matching processing using edge information, for example, presents a problem in that the matching processing takes much time or matching accuracy is not improved. In order to solve such a problem, an image processing device removes noise by smoothing processing using for example an averaging filter or a Gaussian filter. However, when noise is removed by the smoothing processing, edge parts are blurred more, and it becomes difficult to extract edge information.
Accordingly, an image processing device performs filter processing using a median filter, thereby removing noise in the form of dots without losing the sharpness of edge parts. In addition, Japanese Patent Laid-Open No. 2006-014024 (hereinafter referred to as Patent Document 1) discloses smoothing processing performed using only pixels whose differences in luminance value are within a predetermined threshold value, thereby enabling the reduction of false contours and noise without blurring contours while retaining the contours of an input signal.
In addition, when noise in a large region is desired to be removed, filter operation in a wide range is necessary. In this case, simply widening a filter range invites an increase in the scale of an arithmetic circuit and processing time. In order to deal with this, Japanese Patent Laid-Open No. 2010-157163 discloses performing filter processing after reducing an input image.